Data measurement and processing of the hottest sys

  • Detail

MTF data measurement and processing of the system

I. blade scanning

when scanning the blade, it is required that the scanning optical hole must fully reflect the density distribution law or light intensity change law of the linear image. The above requirements can be met by using the slit optical hole scanning blade:

if the slit s (x) with the slit width of D and the slit height of H is used as the scanning optical hole, when H D, the slit height is greater than the slit width as much as possible, but less than the length of the scanning blade; When the slit width is as small as possible and at least less than the line width of the system's limit resolution, the density distribution law of the linear image can be fully reflected. Because for the slit with width D, its slit function is

and its spectral function takes s (0) =d=1 when f=0. After domestication:

from s in Figure 7.5( ξ) And S (f), in the interval of, s (f) is negative. This is because the slit width causes the black-and-white inversion of the figure. Therefore, in the inter purchase measurement, if the slit width D meets the maximum effective frequency component F of the system signal, the black-and-white inversion can be avoided. The image produced in the electronic plate making system is mainly used for visual interpretation. If it can meet (R is the resolution of the system), that is, the seam width is not greater than the line width of the limit resolution, the phenomenon of black-and-white inversion can be eliminated. Therefore, the measurement slit width is selected as 2.5 × one hundred μ OK. When the scan interval is Δ X hour

according to the discrete density points obtained by sampling, the change law of blade density function d (x) can be correctly expressed, and the spectrum can be determined from the discrete density points

press Δ There are two scanning methods for the relationship between X and D:

① normal sampling:

② overlapping sampling:

because the blade density function is a very complex function, it is very difficult to determine its f in advance, and there are many kinds of electronic color separators in use, and the cut-off frequency is not the same. According to relevant data, the maximum cut-off frequency. Calculated. The actual sampling interval is

second, the calculation of line spread function L (x)

in order to obtain L (x), we must first carry out photosensitive determination, and calculate h=f[d (x)]. It can be seen from the principle of MTF that the linear diffusion function L (x) of the system is the first derivative of the knife function material, which also needs to resist the corrosion of cutting fluid L (x):

if the Fourier transform is performed on L and the other is hardener (x), MTF can be obtained. The light intensity distribution I (x) at any point X of the blade curve shown in Figure 7.6 is the sum of the line spread function values corresponding to the image at each point:

in addition, the optical wedge test piece of the image negative must be placed on the PDS micrometer densitometer, and the blade scanning must be carried out under the same conditions, so as to establish the mathematical model of h=f[d (x)] and calculate H (x)

III. preprocessing of scanning data

all scanning is carried out on PDS micrometer densitometer. In order to obtain accurate data, each blade is scanned continuously for 15 pieces (as shown in Figure 7.7), and then its average value is taken as the data for calculating the blade function. At the same time, the photoelectric noise and emulsion particle noise of PDS in the scanning process are reduced:

it is difficult to find the ideal straight edge when selecting the blade from the actual image. The following situations often occur:

① the length of the straight edge of the image is limited

② the straight edge of the image is not "absolutely" straight

therefore, there is a large deviation in some conditions of the data obtained by scanning, as shown in Figure 7.8 if the traditional requirement to reach a specific zinc content is still taken as the standard. This requires manual intervention. Check the original data, remove the items with large deviation, and then take the average value

IV. filtering processing of D (x)

in the sampling process of any image signal, the measured data will inevitably have errors. The image signal always has an upper limit cut-off frequency, and the signal whose sampling frequency is always greater than that near f is mainly noise signal. In this way, the signals obtained after sampling contain noise, which is represented by high-frequency components

the D (x) curve obtained by PDS jitters violently due to the influence of scanner photoelectric noise and photosensitive emulsion particle noise. Although the average value is preprocessed, the jitter is still obvious, so it is necessary to adopt sum filtering method to process d (x) before data operation, so as to eliminate the influence of noise

filtering methods mainly include spatial domain filtering method and spectral domain filtering method. In the process of filtering, the spatial domain filtering method mainly adopts the local filtering method

the so-called local filtering method is also called curve moving smoothing method. It is based on the curve fitting principle of polynomial least square method. It is a moving smoothing filtering method with an nth degree polynomial function as the filtering operator. For example, the five point cubic moving smoothing formula is:

in the calculation of MTF, the local smoothing derivation method is mainly applied after spectrum filtering. At this time, the noise is not strong because the signal has been filtered several times. This method can obtain satisfactory results, and the algorithm is simple and fast. When calculating L (x), it is proved that the five point cubic derivative filtering method is the best

spectral domain filtering method is to attenuate high-frequency components and make low-frequency components pass smoothly. It is usually difficult to determine the cut-off frequency f when filtering in the spectral domain. Under the influence of noise n (T), it is usually difficult to determine the cut-off frequency f when filtering the spectral amplitude of X (T). Under the influence of noise n (T), the spectral amplitude x (f) of X (T) decreases to a certain extent, and then jitters around a certain value. At this time, it can be considered that the spectral amplitude x (f) of X (T) is mainly determined by noise; It can also be considered that the signal at this time is noise. In the experiment, this method is used to filter and determine F, as shown in Figure 7.9

how many high-frequency components need to be filtered in order to maintain effective signals as much as possible and suppress noise appropriately is also a difficult problem to solve

in spatial filtering method, the filtering times are usually limited by the starting point of the blade. When the starting point changes during filtering, it is difficult to determine the number of filtering times. Using spectral domain filtering can solve the problem of the beginning and end point change in spatial domain filtering. As long as the cut-off frequency f is properly selected, the noise may be effectively suppressed after one filtering. As for the determination of F, the spectrum of the discrete sampling signal of the blade can be obtained by fast Fourier transform (FFT), and then the cut-off frequency f can be obtained by analyzing the spectrum, and then a filter and inverse Fourier transform can be carried out to recover the filtered signal. In this way, the noise can be effectively suppressed without distorting the signal; It not only ensures the characteristics of the signal, but also reduces the data processing time. Figure 7.10 shows the comparison of the effects of the same data after spatial domain filtering method and spectral domain wave method respectively

Copyright © 2011 JIN SHI